Created attachment 796914 [details] packstack installation file Description of problem: cannot configure openstack AIO to work with remote tgt storage. cinder doesn't trey to connect to the remote server although the iscsi target directs to the tgt server. 2013-09-12 17:26:35.493 9779 ERROR cinder.brick.local_dev.lvm [req-af7e889e-5989-4f69-90eb-090f96bdd0ea None None] Unable to locate Volume Group cinder-volumes 2013-09-12 17:26:35.494 9779 ERROR cinder.service [req-af7e889e-5989-4f69-90eb-090f96bdd0ea None None] Unhandled exception Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. Install all the openstack services except cinder with packstack. 2. Install the Cinder manually and configure it according to: http://docs.openstack.org/grizzly/openstack-compute/install/apt/content/cinder-install.html 3. Install and configure a remote TGT server and configure the storage setup according to the same manual. Actual results: Error: cinder can't locate the cinder-volumes Expected results: an AIO openstack server manages its volumes remotely. Additional info: 2013-09-12 17:26:34.711 9766 INFO cinder.service [-] Starting 1 workers 2013-09-12 17:26:34.713 9766 INFO cinder.service [-] Started child 9779 2013-09-12 17:26:34.716 9779 AUDIT cinder.service [-] Starting cinder-volume node (version 2013.2) 2013-09-12 17:26:35.086 9779 INFO cinder.openstack.common.rpc.impl_qpid [req-af7e889e-5989-4f69-90eb-090f96bdd0ea None None] Connected to AMQP server on 10.35.104.16:5672 2013-09-12 17:26:35.107 9779 INFO cinder.volume.manager [req-af7e889e-5989-4f69-90eb-090f96bdd0ea None None] Starting volume driver LVMISCSIDriver (2.0.0) 2013-09-12 17:26:35.493 9779 ERROR cinder.brick.local_dev.lvm [req-af7e889e-5989-4f69-90eb-090f96bdd0ea None None] Unable to locate Volume Group cinder-volumes 2013-09-12 17:26:35.494 9779 ERROR cinder.service [req-af7e889e-5989-4f69-90eb-090f96bdd0ea None None] Unhandled exception 2013-09-12 17:26:35.494 9779 TRACE cinder.service Traceback (most recent call last): 2013-09-12 17:26:35.494 9779 TRACE cinder.service File "/usr/lib/python2.6/site-packages/cinder/service.py", line 228, in _start_child 2013-09-12 17:26:35.494 9779 TRACE cinder.service self._child_process(wrap.server) 2013-09-12 17:26:35.494 9779 TRACE cinder.service File "/usr/lib/python2.6/site-packages/cinder/service.py", line 205, in _child_process 2013-09-12 17:26:35.494 9779 TRACE cinder.service launcher.run_server(server) 2013-09-12 17:26:35.494 9779 TRACE cinder.service File "/usr/lib/python2.6/site-packages/cinder/service.py", line 96, in run_server 2013-09-12 17:26:35.494 9779 TRACE cinder.service server.start() 2013-09-12 17:26:35.494 9779 TRACE cinder.service File "/usr/lib/python2.6/site-packages/cinder/service.py", line 385, in start 2013-09-12 17:26:35.494 9779 TRACE cinder.service self.manager.init_host() 2013-09-12 17:26:35.494 9779 TRACE cinder.service File "/usr/lib/python2.6/site-packages/cinder/volume/manager.py", line 170, in init_host 2013-09-12 17:26:35.494 9779 TRACE cinder.service self.driver.check_for_setup_error() 2013-09-12 17:26:35.494 9779 TRACE cinder.service File "/usr/lib/python2.6/site-packages/cinder/volume/drivers/lvm.py", line 92, in check_for_setup_error 2013-09-12 17:26:35.494 9779 TRACE cinder.service raise exception.VolumeBackendAPIException(data=message) 2013-09-12 17:26:35.494 9779 TRACE cinder.service VolumeBackendAPIException: Bad or unexpected response from the storage volume backend API: Volume Group cinder-volumes does not exist 2013-09-12 17:26:35.494 9779 TRACE cinder.service 2013-09-12 17:26:35.508 9766 INFO cinder.service [-] Child 9779 exited with status 2 2013-09-12 17:26:35.508 9766 INFO cinder.service [-] _wait_child 1 2013-09-12 17:26:35.509 9766 INFO cinder.service [-] wait wrap.failed True
Created attachment 796915 [details] Cinder log files from /var/log/cinder
The error here indicates that there is no cinder-volumes VG on the node running the Cinder volume service. I'm not sure I understand what you mean by "remote TGT". The VG and cinder-volume service (which manages tgtd) should be deployed on the same node, is that the case here?
Remote TGT means that a TGT storage device can be managed by the Cinder while it is on a different machine. The problem is that the Cinder doesn't have a driver to support this scenario, as for an example, a driver for gluster allows the Cinder to create/delete/attach/etc. volumes on a remote storage device. Though the cinder.conf configuration file points the iscsi address to the TGT storage device (that is on a different machine) the Cinder searches for the cinder-volumes localy.
(In reply to Yogev Rabl from comment #4) The iscsi_ip_address option is intended to let you pick the intended IP to use on a multi-homed box, where tgtd is still local and managed by Cinder. As you note, Cinder doesn't have a driver which supports this scenario. I think the best thing to fit what you are requesting is the driver work which is planned for Icehouse that will enable use of targetd. The targetd service manages LVM and LIO to export volumes over iSCSI from a remote box and seems to be what you are after here.
I'd suggest we mark this as a duplicate of bug 890202.
tgtd is being replaced with lio and targetd is the way to manage lio remotely so Bug 890202 will in effect supply the needed functionality here. *** This bug has been marked as a duplicate of bug 890202 ***